Image Classification on CIFAR100 12/255 noise (test)
53.46Natural AccuracyPGD-2
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PGD-2Backbone=PreActResNet-182024.04 | 53.46 | 17.16 | |
| ZeroGradBackbone=PreActResNet-182024.04 | 49.07 | 14.76 | |
| PGD-10(20)Backbone=PreActResNet-182024.04 | 47.42 | 20.29 | |
| MultiGradBackbone=PreActResNet-182024.04 | 46.81 | 16.56 | |
| N-FGSMBackbone=PreActResNet-182024.04 | 46.16 | 18.75 | |
| N-AAERBackbone=PreActResNet-182024.04 | 45.98 | 19.03 | |
| Grad AlignBackbone=PreActResNet-182024.04 | 45.83 | 18.13 | |
| RS-AAERBackbone=PreActResNet-182024.04 | 44.06 | 16.41 | |
| FreeATBackbone=PreActResNet-182024.04 | 39.84 | 16.57 | |
| RS-FGSMBackbone=PreActResNet-182024.04 | 18.19 | 0 |